QuestBench - Mean Score: leaderboard
Metric: Mean score (0-100) against question-specific grading criteria over the 256 Chinese deep research questions, with identical web search and visit tools and a nominal 50-call budget per question; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 67.12 |
| 2 | Claude Opus 4.7 | 57.79 |
| 3 | Gemini 3.1 Pro (Preview) | 43.07 |
| 4 | GLM-5.1 | 42.3 |
| 5 | DeepSeek V4 Pro | 38.24 |
| 6 | Kimi K2.6 | 37.74 |
| 7 | Qwen 3.6 Plus | 36.05 |
| 8 | MiMo-V2.5-Pro | 32.05 |
| 9 | Kimi K2.5 | 30.48 |
| 10 | MiniMax-M2.7 | 29.29 |
| 11 | DeepSeek V3.2 | 14.58 |
Interactive version: theaggregate.ai/benchmark?slug=questbench-mean-score · How It Works · Data refreshed daily, snapshot 2026-10-07.